A critical function of the vertebrate retina is to change its sensitivity based on the recent history of the stimulus in order to maintain a visual response when the environment changes. This process, known as adaptation, occurs in multiple forms, although many aspects of the cellular and circuit origin of these computations remain unknown. Recently, it was found that certain retinal ganglion cells increase their sensitivity following a strong stimulus. These sensitizing ganglion cells maintain a high sensitivity to weak stimuli, even when other types of ganglion cells adapt and fall below threshold. This proposal aims to analyze the circuit basis for the adaptive computation of sensitization. To understand sensitization arises, we have developed a novel approach to directly measure the functional role of individual retinal interneurons. We record the intracellular visual response from a single interneuron, while simultaneously recording from a population of retinal ganglion cells using a multielectrode array. Then, by playing back an altered version of the cell's own signal using injected current, we directly probe how that cell's output changes the behavior of the circuit. Mathematical models are used to explain how the responses of interneurons combine together to yield the responses of ganglion cells. Amacrine cells are a diverse population of inhibitory interneurons in the retina, most with unknown function. Some amacrine cells are known to adapt to the contrast of the stimulus, but the functional role of this process is unknown. The goal of this proposal is test the hypothesis that adaptive inhibition generates sensitization, and to develop a quantitative circuit level description of how sensitization arises. The specific goals are: 1) In the salamander and mouse retina, we will measure the spatiotemporal structure of how adaptive excitation and inhibition combine to generate sensitization, and capture these properties with a computational model. Using pharmacology, we will identify the broad class of amacrine cells that are essential to sensitization. 2) Using intracellular and multielectrode recording, we will measure the computation of sensitization through salamander retinal circuitry by recording from interneurons during sensitization, and then directly measure their connectivity to different classes of ganglion cells. Synaptic currents in salamander and mouse ganglion cells will be analyzed using whole cell recordings. 3) We will directly measure how changes in transmission from single inhibitory amacrine cells generate sensitization in the intact salamander retina. Understanding how a diverse population of neurons combines to perform neural functions is a critical barrier to our understanding of retinal mechanisms and diseases involving the degeneration of the retinal circuitry. These findings will be essential to understanding basic mechanisms of how retinal circuitry processes information and will be useful in the design of electronic retinal prosthesis systems.